An Analysis of the AIG Case: Understanding Systemic Risk and Its Relation to Insurance

An Analysis of the AIG Case: Understanding Systemic Risk and Its Relation to Insurance Dr. Etti Baranoff1

Abstract This study describes the AIG2 model of operations prior to the conglomerate failure up to the point when the liquidity crisis triggered the massive bailout by the U.S. government. It is a study designed to provide understanding of the key factors in the demise of AIG in relationship to systemic risks in insurance. The main contribution of this report is the delineation of the key internal factors from the external macro market and regulatory factors that contributed to the failure. We regard the latter as macro factors underpinning the foundation that propelled the activities of AIG Financial Products Unit (AIGFP). The study shows that if it were not for the “non-insurance” activities of the AIGFP under the AIG holding company, the averted collapse (with the bailout), in all likelihood, would have been avoided. The main key takeaways are: AIGFP was not an insurance company; AIGFP was not regulated by state-based insurance regulations; and AIGFP’s credit default swaps were the key factor to the AIG collapse. As global regulators look into indicators for systemically important financial institutions (SIFIs), the following macro factors should be integrated into any newly created regulatory framework: 1) use credit ratings with care and avoid exploitation of high ratings; 2) be aware of banks’ capital being replaced by new opaque financial products; 3) remove gaps in regulations and require transparency; 4) forbid companies to select their own regulatory bodies; 5) understand insurance vs. noninsurance or quasi-banking activities and products; and 6) create clarity to 1. Associate Professor, Insurance and Finance, Virginia Commonwealth University, Richmond, VA. Research Director, Insurance and Finance, The Geneva Association. 2. The term “AIG” refers to the total scope of consolidation of the American International Group, Inc. (AIG) holding company and does not mean specific legal entities, but otherwise stated. © 2012 National Association of Insurance Commissioners

Journal of Insurance Regulation

244

delineate between the banking and insurance models. In brief, the key lesson is that when non-insurance or quasi-banking operations enter the insurance arena, expert insurance supervision is needed to close gaps in regulation. Quoted from the opening paragraph of the chapter “September 2008: The Bailout of AIG,” of the January 2011 The Financial Crisis Inquiry Report of the National Commission on the Causes of the Financial and Economic Crisis in the United States: “The AIG corporate empire held more than $1 trillion in assets, but most of the liquid assets, including cash, were held by regulated insurance subsidiaries whose regulators did not allow the cash to flow freely up to the holding company, much less out to troubled subsidiaries such as AIG Financial Products.”

I.

Introduction

The objective of this article is to identify links between the AIG failure and the understanding of systemic risks at financial institutions and to identify lessons to be drawn for safeguarding the financial stability of the insurance sector. Using the chronicle of the near collapse of AIG and its subsequent bailout with $182 billion3 in government funds, this study uncovers internal and external (macro) factors explaining the collapse. The internal factors are those activities that were generated from within AIG and are also called “inside” activities. The macro factors are those external to AIG, but important as drivers to their behavior. Without some of the external underpinnings, their operations would not have been allowed and the failure might have been averted. Thus, both the inside and external macro factors intertwine to explain the causes for the AIG debacle and, subsequently, the understanding of systemic risk determinants to create a model for systemic risks in insurance.4 Prior to 2008, AIG operated a successful global insurance business that provided the basis for the company’s stellar credit ratings. Utilizing their good standing and broad-based global reach, AIG permitted financial innovators to create new financial products operating outside the purview of most regulatory surveillance bodies. To be able to operate outside of the insurance arena and its 200-year-old regulatory structure in the United States, AIG created the AIG Financial Products Corp. (AIGFP) to manage these new financial products under its holding company.

3. Orol (2010). 4. This delineation may appear artificial, but it is designed to pinpoint areas that can generate systemic risks and clarify the chronicle of the demise in terms of its components. © 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

245

The holding company was regulated by the federal Office of Thrift Supervision (OTS), not the state-based insurance regulators. OTS regulators admitted their inability and incapacity to regulate a sophisticated unit such as the AIGFP.5 AIGFP sold credit default swaps (CDS), an unregulated product which was at the core of the liquidity crisis that brought the company down. Other cited causes for the trouble—securities lending activities and investments in mortgage-backed securities (MBS)—exacerbated the liquidity shortage. Those two factors on their own, without the CDS, probably would not have led to AIG’s demise. Even though AIG was aggressive in pursuing the two activities, the calls to post-cash collateral for the CDS were key to the company’s downfall when the housing markets collapsed in 2008 and AIG’s ratings were downgraded. These cash calls illuminate the faulty CDS contract design and the glaring lack of supervision of this product. The calls for cash collateral depleted the liquidity of AIG, its reputation and the trust of all counterparties. The liquidity crisis was averted by the bailout from the U.S. government. The main contribution of this report is the delineation of the key internal factors from the external macro market and regulatory factors that contributed to the failure. We regard the latter as macro factors underpinning the foundation that propelled the activities of AIGFP. As global regulators look into indicators for systemically important financial institutions (SIFIs), these macro factors should be integrated into any newly created regulatory framework. The inside and external (macro) factors that led to AIG’s collapse are summarized in Table 1.

5. The Financial Crisis Inquiry Report, 2011. National Commission on the Causes of the Financial and Economic Crisis in the United States. © 2012 National Association of Insurance Commissioners

Journal of Insurance Regulation

246

Table 1: The inside and external (macro) factors that led to AIG’s near collapse and bailout External (Macro) Factors to the Collapse Reliance on rating agencies led to faulty “trust” in the markets.

Inside Factors to the Collapse =>

Housing market bubble (entitlement ideology) led to subprime mortgages and growth in mortgage-backed securities (MBS) (securitization) of toxic loans.

=>

Lax banking and OTS regulation led banks to use CDS from highly rated providers. There was no derivatives regulation (the free markets ideology). State insurance regulators were not part of the oversight structure of the non-insurance operations of the AIG holding company and AIGFP.

=>

AIG’s strong insurance operation provided a stellar rating. The high rating led highly sophisticated financial innovators to create the AIG Financial Products Corp. (AIGFP) under the AIG holding company. AIGFP sold credit default swaps (CDS) and other derivatives. AIGFP responded to the increased needs for CDS to provide “apparent assurances.” The CDS became enablers to growth in securitization and leveraging by banks. The circular motion led AIGFP to cover subprime mortgages despite their stricter guidelines. When the housing bubble burst, AIGFP was the holder of the “apparent safety net” to many banks. The liquidity crisis erupted. AIGFP exposure in CDS grew to more than $500 billion by 2008. There were no checks and balances over the operations of the unit. OTS regulators noted their lack of expertise. Without derivatives regulations, there were no transparencies regarding the CDS. CDS contracts had faulty designs with fast-paced cash collateral calls for downgrades of AIG. The liquidity crisis of AIGFP was not transparent to the whole conglomerate in time. AIG’s faulty financial models neglected to account for some important assumptions. CDS were (and are not) insurance contracts and, therefore, lacked the “safety valves” of insurance contracts.

© 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

247

Table 1 continued

Insurance regulation has been strong over the AIG insurance units and insurance products.

=>

Financial markets crisis erupted and Lehman Brothers was allowed to collapse. The U.S. government recognized the interconnectedness of the AIGFP unit’s activities worldwide.

=>

The securities lending activities of AIG insurance units and the investments in MBS were more aggressive than the regulatory guidelines. AIG had to provide cash pools to adhere to state-based insurance regulations. As such, the insurance subsidiaries did not sell the toxic assets just acquired. These activities continued simultaneously as AIGFP was running out of cash, thus exasperating the liquidity crisis. State insurance regulators did not allow the use of the $1 trillion in the insurance companies’ assets to financially assist AIGFP, a non-insurance entity. Efforts to find capital market solutions failed and the U.S. government provided AIG with a $182 billion bailout.

This article shows that when systemic risk enters into the discussions of AIG, it is key to understand that: 1. 2. 3.

AIGFP was not an insurance company. AIGFP was not regulated by state-based insurance regulations. AIGFP’s credit default swaps were the key factor to the AIG collapse.

In addition, there are major macroprudential lessons to be learned from the AIG failure, as follows: 1. 2. 3. 4. 5. 6.

Use credit ratings with care: Do not allow exploitation of high ratings. Be aware of banks’ capital being replaced by new opaque financial products. Remove gaps in regulations and require transparency. Forbid companies to select their own regulatory bodies. Understand insurance vs. non-insurance or quasi-banking activities and products. Create clarity to delineate between the banking and insurance models.

© 2012 National Association of Insurance Commissioners

248

Journal of Insurance Regulation

Using many sources, this article shows how the inside activities and macro factors led to AIG’s demise and, ultimately, the necessary governmental rescue actions. Section II provides an overview of the AIG frameworks and creates models to depict the interactions between the forces that brought about to the calamity of the AIGFP unit. Section III is a brief discussion of the definition for systemic risk in connection to the AIG story, an explanation of the differences between insurance and CDS, and the nature of securities lending. This section also provides a brief overview of the nature of insurance regulation. Section IV focuses on AIG’s inside factors, while section V focuses on the external macro factors. Section VI brings both factors together and follows the sequence to the collapse as shown in Figure 4. The report concludes with a summary and lessons to be learned to avoid future calamities.

II. Overview and the AIG Framework in Models As will be shown later, and is fully supported by the recently published 2011 report of the National Commission on the Causes of the Financial and Economic Crisis in the United States, the main internal factor to the AIG collapse was the credit default swaps6 (CDS) sold by the AIG Financial Products Corp. (AIGFP): The Commission concludes AIG failed and was rescued by the government primarily because its enormous sales of credit default swaps were made without putting up initial collateral, setting aside capital reserves, or hedging its exposure—a profound failure in corporate governance, particularly its risk management practices.7 Harrington (2009) and Sjostrum (2009), as well as most of the reports about the AIG failure, also include the “securities lending activities” and investment in mortgage-backed securities (MBS) as causes of the failure. We explain in this report why the CDS exposure was the key ingredient to the failure. While AIG’s securities lending activities were aggressive relative to the New York state insurance regulatory requirements at the time, and led to a need to put cash

6. A credit default swap (CDS) is an agreement in which the protection buyer of the CDS makes a series of payments (often referred to as the CDS “fee” or “spread”) to the protection seller and, in exchange, receives a payoff if a credit instrument (typically a bond or loan) experiences a credit event. 7. See National Commission on the Causes of the Financial and Economic Crisis in the United States (2011), p. 352. © 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

249

collateral in pools,8 this activity on its own would have not taken AIG down9 without the liquidity crisis generated by the exposure of the CDS and the call for cash collateral upon each rating downgrade of the conglomerate. As activities regulated by insurance regulators, the collateral requirements were set to avoid potential solvency concerns. The cash needs of this activity were not in crisis, but were aggravated by the liquidity crisis of the CDS exposure. The structure of AIG within regulatory frameworks is shown in Figure 1. The AIGFP unit was part of the AIG holding company structure, but it was legally and operationally separated from the insurance operations. The insurance operations had been under the scrutiny of the U.S. insurance regulators and generated a tripleA credit rating, highly coveted among financial institutions. The other side of the structure was under permissive banking regulation and “no regulation” for derivatives and CDS. Based on the insurance success and the high rating, AIGFP, a non-insurance entity, sold “$2.7 trillion worth of swap contracts and positions; 50,000 outstanding trades; 2,000 firms involved on the other side of those trades; and (had) 450 employees in six offices around the world.”10 The CDS pricing and promises were developed based on sophisticated models and assumptions. The underlying assumptions were based on two crucial factors: 1) AIG could keep its

8. As gleaned from the 2008 AIG 10K report, p. 6, “Continuing Liquidity Pressures”: “Historically, under AIG’s securities lending program, cash collateral was received from borrowers and invested by AIG primarily in fixed maturity securities to earn a spread. AIG had received cash collateral from borrowers of 100 to 102 percent of the value of the loaned securities. In light of more favorable terms offered by other lenders of securities, AIG accepted cash advanced by borrowers of less than the 102 percent historically required by insurance regulators. Under an agreement with its insurance company subsidiaries participating in the securities lending program, AIG parent deposited collateral in an amount sufficient to address the deficit. AIG parent also deposited amounts into the collateral pool to offset losses realized by the pool in connection with sales of impaired securities. Aggregate deposits by AIG parent to or for the benefit of the securities lending collateral pool through August 31, 2008 totaled $3.3 billion.” 9. Per discussion with experts in the field of securities lending, we learned that this activity is in essence an investment activity facilitated by leverage. The lender of the securities, the insurer, invests the cash collateral pledged by the borrower. If the investments turn sour, the lender would incur an unrealized loss, which would be realized upon a default of the investment or if they chose to sell the investment — as would be the case with any other investment held by the insurer. The lender has an unconditional obligation to the borrower to return the cash collateral upon demand. And, because collateral is marked to market daily, presumably the borrower would not suffer a loss, even if an insurer becomes insolvent and is unable or chooses not to return the cash to the borrower. The borrower under such circumstances could/would sell the borrowed securities and close out the transactions. The caveat to all this is that under some state laws, regulators could invoke a stay and possibly prevent the borrower from closing out the loan. But even if that were to occur, the borrower still has the collateral, which they presumably could pledge to others if they needed liquidity. Thus, could insolvency cause a systemic risk? It does not appear to be the case here. 10. O’Harrow Jr. and Dennis (2008). © 2012 National Association of Insurance Commissioners

Journal of Insurance Regulation

250

high rating;11 and 2) CDS were to cover high-quality debt instruments with minimal potential of non-performance. Based on these assumptions, embedded in the sophisticated models created by the ingenious employees of AIGFP, CDS contracts included provisions requiring posting of cash collateral in case AIG’s credit rating would decline. Thus, the mechanisms in the embedded models and contracts for CDS were dependent on no or minimal defaults in debt and mortgage loans for housing, along with keeping AIG’s high credit ratings (the pink arrow in Figure 1) generated by the strong insurance operations (an important factor).

Figure 1: AIG Structure Relative to Regulation (Blue are internal AIG factors) Lax Banking /Thrift Regulation/No Derivatives Regulation

Insurance Regulation AIG Holdings Triple-A Rating AIGFP AIG Insurance Operations

Financial Models and Innovation – Credit Default Swaps

11. As gleaned from the 2008 AIG 10K report, p. 42, downgrades and posting of cash collateral went hand-in-hand: “In the event of a further downgrade of AIG’s long-term senior debt ratings, AIGFP would be required to post additional collateral and AIG or certain of AIGFP’s counterparties would be permitted to elect early termination of contracts. It is estimated that as of the close of business on February 18, 2009, based on AIGFP’s outstanding municipal GIAs, secured funding arrangements and financial derivative transactions (including AIGFP’s super senior credit default swap portfolio) at that date, a one-notch downgrade of AIG’s longterm senior debt ratings to Baa by Moody’s and BBB+ by S&P would permit counterparties to make additional collateral calls and permit either AIGFP or the counterparties to elect early termination of contracts, resulting in up to approximately $8 billion of corresponding collateral postings and termination payments, a two-notch downgrade to Baa by Moody’s and BBB by S&P would result in approximately $2 billion in additional collateral postings and termination payments, and a three-notch downgrade to Baa by Moody’s and BBB by S&P would result in approximately $1 billion in additional collateral and termination payments.”

© 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

251

The main ideologies underpinning the huge growth in AIG’s CDS exposure were the “free market” philosophy and the belief that “everyone deserves to own a home.” These credos manifested themselves in lenient banking regulation and no regulation for derivatives instruments. The innovative financial instruments such as CDS, therefore, were allowed to grow without the necessary checks and balances. The boom in the U.S. housing market was propelled by the securitization of MBS, along with a high level of leveraging and slack underwriting. The MBS included too many low-grade subprime mortgages. Without surveillance and lack of risk management in the mortgage markets, along with dependency on the high rating of the AIG conglomerate, AIGFP was allowed to generate an enormous exposure in CDS ($533 billion in notional amount at the end of 200712). The January 2011 National Commission on the Causes of the Financial and Economic Crisis in the United States’ Report proceeds as follows: AIG’s failure was possible because of the sweeping deregulation of over-the counter (OTC) derivatives, including credit default swaps, which effectively eliminated federal and state regulation of these products, including capital and margin requirements that would have lessened the likelihood of AIG’s failure. The OTC derivatives market’s lack of transparency and of effective price discovery exacerbated the collateral disputes of AIG and Goldman Sachs and similar disputes between other derivatives counterparties. As shown in Figure 2, there was a circular motion with interdependencies among the housing markets, the bundled debt securities and the CDS. As more CDS provided “apparent security,” more debt was allowed, and in this circular motion, more mortgages were issued with erosion in underwriting standards, and with more bundling of the debt. The stated political aim of the U.S. government that “everyone deserves to own a home” increased the number of subprime mortgages and lowered the quality of the debt secured by AIGFP.

12. See Harrington (2009), p. 790. © 2012 National Association of Insurance Commissioners

Journal of Insurance Regulation

252

Figure 2: The circular motion of the housing boom, the bundled securities and AIG credit default swaps

Housing Market

Bundling of mortgages AIG Credit Default Swaps-huge exposure

AIG was operating with highly risky CDS within markets and regulatory frameworks that, at any moment, could be triggered into a massive systemic collapse. The structure is depicted in Figure 3, which combines Figure 1 and Figure 2. Figure 3 (combining Figures 1 & 2): AIG Structure Relative to Regulation, Housing, Mortgage Backed Securities and Credit Default Swaps – Before the Collapse Lax Banking /Thrift Regulation/No Derivatives Regulation

Insurance Regulation AIG Holdings Triple-A Rating

Housing Market

AIGFP Financial Models and Innovation – Credit Default Swaps**

AIG Insurance Operations

Bundling of Mortgages Credit Default Swaps

** Huge Exposure of Credit Default Swaps

© 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

253

When the housing bubble burst and the credit rating organizations lowered AIG’s rating, the sheer size of AIG’s CDS exposure led to the collapse as shown in Figure 4. Under the CDS contracts, without the high rating, AIG had to post cash collateral for the credit instruments they covered with the CDS. Liquidity evaporated and the bailout was perceived to be 13 the only solution to avoid the potential for a complete destabilization of the entire financial system.

Figure 4: After Ratings are Lowered -- AIG Collapse (Bailout) Lax Banking /Thrift Regulation/No Derivatives Regulation

Insurance Regulation AIG Holdings

AIG Insurance Operations is Solid ($1 trillion in assets) – All policyholders are paid

AIGFP Financial Models and Innovation – Credit Default Swaps**

Liquidity Disaster

Housing Market Collapse

Bundling of Mortgages AIG Credit Default Swaps

** Huge Exposure of Credit Default Swaps requiring liquidity

Thus, the principle finding in this article is that unexpected and unmodeled external macro risks suddenly exposed the risky activities of AIGFP. The strength of the insurance operation led to adopting non-insurance financial instruments outside of the insurance operation. These products, generated by AIGFP, depended solely on the strength of the insurance business. Without the external macro circumstances of lax and/or no regulation in key areas (i.e., the financial realm, not 13. AIG, in its report to the U.S. Treasury Department, portrayed the situation as dire because of “run on the bank” for life insurance. See also Sorkin (2009). There are a few arguments refuting AIG’s assertions. The report wrongfully regards life insurance policyholders as uninsurable people with inability to obtain new policies from the competition. The life insurance industry is known for its competitiveness. Many insurers were ready to accept the AIG policyholders in case AIG became insolvent. Because resolution of insurance companies is consistently an orderly process that includes state guaranty funds, many oppose AIG’s report about a potential “run on the bank.” This unfortunately instilled misperceptions about the nature of the insurance model as opposed to the banking model. © 2012 National Association of Insurance Commissioners

254

Journal of Insurance Regulation

insurance), the growth in the CDS exposure of AIGFP would have lacked a key enabling element. Supporting this conclusion is the following finding by The Financial Crisis Inquiry Report: AIG engaged in regulatory arbitrage by setting up a major business in this unregulated product, locating much of the business in London, and selecting a weak federal regulator, the Office of Thrift Supervision (OTS). The OTS failed to effectively exercise its authority over AIG and its affiliates: it lacked the capability to supervise an institution of the size and complexity of AIG, did not recognize the risks inherent in AIG’s sales of credit default swaps, and did not understand its responsibility to oversee the entire company, including AIG Financial Products. Furthermore, because of the deregulation of OTC derivatives, state insurance supervisors were barred from regulating AIG’s sale of credit default swaps even though they were similar in effect to insurance contracts. If they had been regulated as insurance contracts, AIG would have been required to maintain adequate capital reserves, would not have been able to enter into contracts requiring the posting of collateral, and would not have been able to provide default protection to speculators; thus AIG would have been prevented from acting in such a risky manner. If we play the “what if” game by taking each of the external macro factors out, we can show how the absence of any enabling factor could have stopped the sequence of events that led to the disaster shown in Figure 4.14 Sorkin (2009) describes the situation as follows: “AIG used its triple-A rating from the insurance part of its business to run a huge casino that then overwhelmed the entire business.” Edward M. Liddy, who became AIG’s chief executive after the bailout added: “It’s an interesting structure where you have an insurance company that works really well and on top of it is a holding company and the holding company’s biggest asset is this huge hedge fund.” Ron Shelp, the author of Fallen Giant15 described AIG as a “risk-taker” during his Bloomberg interview.16 Our understanding here is that the AIGFP unit took advantage of every crack in the regulatory structure. Such arbitrage risk should be considered very closely when developing a future regulatory oversight structure that could prevent systemic risk from collapsing the system. Considering exclusively the inside or internal elements of systemically risky activities without examining the contributory 14. This is shown in the appendix section of the report. 15. Fallen Giant: The Amazing Story of Hank Greenberg and the History of AIG tells the story of global insurance giant AIG, from its first business transaction in China in 1919 to the exit of its longtime chairman and CEO Maurice “Hank” Greenberg in 2005. 16. www.bloomberg.com/video/61196958. © 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

255

external macro elements leaves a major gap in the regulatory surveillance for systemic risks and opens the door to new “creative” activities that could lead to new crises in the future. Finally, the U.S. National Commission on the Causes of the Financial and Economic Crisis in the United States adds: AIG was so interconnected with many large commercial banks, investment banks, and other financial institutions through counterparty credit relationships on credit default swaps and other activities such as securities lending that its potential failure created systemic risk. The government concluded AIG was too big to fail and committed more than $180 billion to its rescue. Without the bailout, AIG’s default and collapse could have brought down its counterparties, causing cascading losses and collapses throughout the financial system.

III. Systemic Risks in Insurance, Insurance vs. CDS, Securities Lending and Insurance Regulation Systemic Risks Because this article is focused on better understanding systemic risks17 in insurance using the AIG debacle as the key case study, we begin with the definitions provided by the Financial Stability Board (FSB) and the International Association of Insurance Supervisors (IAIS) and thoroughly analyzed in The Geneva Association Systemic Risk in Insurance (SRI) reports (2010a and 2010b). The SRI reports employ the FSB criteria (and the IAIS extension of them) for systemic risk using size, interconnectedness, substitutability and timing as the core triggers to examine insurers’ activities for identification of potentially systemically risky activities (pSRA). These activities are regarded as internal to financial institutions. In relationship to the AIGFP demise, all four factors in the definition did play a role. The overall size of the CDS was unprecedented in its exposure. The CDS were completely interconnected to many players in the market and there was no substitutability. The timing was critical and led to the tsunami of the liquidity crunch as shown in Figure 4. Moreover, AIG was one of the conglomerates recognized as “too big to fail.”

17. The lack of definition is thoroughly discussed in The Geneva Association Insurance and Finance Newsletter No. 6, July 2010, “The Lack of an Appropriate Definition of Systemic Risk” by Patrick Liedtke. © 2012 National Association of Insurance Commissioners

256

Journal of Insurance Regulation

Added links to systemic risks are the underpinning external macro factors that led to the collapse. One of the contributions of this article is the observation that the definition of systemic risks cannot ignore the role of “no regulation” or “inadequate regulation” and the importance of rating organizations as key facilitators to systemic failure. When potentially systemic risky behavior is orchestrated because of incentives built into risky markets and gaps in regulatory frameworks, there can be momentum for massive destruction.

Insurance Contracts vs. Credit Default Swaps René M. Stulz (2010) in his paper “Credit Default Swaps and the Credit Crisis”18 provides a well-written comparison between credit default swaps and insurance products that highlights the key difference between them. He notes: …. the parallel between insurance contracts and credit default swaps does not hold in two important ways. First, you do not have to hold the bonds to buy a credit default swap on that bond, whereas with an insurance contract, you typically have to have a direct economic exposure to obtain insurance. Because you don’t have to hold bonds, the amount you insure with a credit default swap is usually called the notional amount. If you buy a credit default swap on Ford for a notional amount of $100 million, you have insurance on $100 million of principal amount of Ford bonds. Second, insurance contracts (mostly) are not traded; in contrast, credit default swap contracts do trade over the counter—that is, a market where traders in different locations communicate and make deals by phone and through electronic messages.” CDS “seem like straightforward financial derivatives that serve standard useful functions: making it easier for credit risks to be borne by those who are in the best position to bear them, enabling financial institutions to make loans they would not otherwise be able to make, and revealing useful information about credit risk in their prices. However, because they covered subprime mortgages, CDS became the culprit that “blew up Wall Street” and caused the demise of AIG. The accounting treatment and the regulatory oversight are very different between CDS and insurance contracts. Insurance contracts are well-structured and have sustained centuries of court cases and refinement, as evident in examining any insurance policy. The insurance regulatory authorities require the insurer to set up a (reserves) liability at policy inception to cover expected losses for every type 18. René M. Stulz is the Reese Chair of Banking and Monetary Economics, The Ohio State University, Columbus, Ohio, and research associate, National Bureau of Economic Research, Cambridge, Mass. © 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

257

of product that is introduced in the marketplace. Against these liabilities, the insurer holds assets to match to the liabilities, a practice known as asset/liability matching. Every asset has risk factors that lead to capital requirements under the risk-based capital (RBC) formula. Such was not the case for the CDS market. As opposed to the requirements imposed on the writer of an insurance contract, the writer of a CDS contract (i.e., the seller of the protection) does not have to post liabilities or reserves as long as cash collateral is not called for. Only if the CDS gets in the money from the buyer’s point of view (i.e., the contract represents an asset to the buyer), does the writer have to set up a liability representing the debt. Also, there was no regulation of the CDS contracts and no oversight. The structure of the CDS contract had not received the historical scrutiny of an insurance contract and the careful regulatory treatment by supervisors. The more than 200-year-old tradition of insurance regulation (discussed below) with its comprehensive codes and regulations was not afforded the opportunity to scrutinize this risky CDS product, which was deemed to be a systemically relevant activity (SRA) in The Geneva Association SRI reports.

Securities Lending Because its securities lending activity was cited as a contributor to the demise of AIG and also noted as a pSRA in The Geneva Association SRI reports, we provide here an explanation of the nature of the securities lending activity.19 Securities lending is rather straightforward. The lending institution transfers securities to the borrower in exchange for cash collateral. The lender pays the borrower a financing or “rebate” rate for the use of the cash collateral. The cash collateral is marked to market daily, such that the value of the collateral posted to the lender always meets or exceeds the value of the loaned securities. The lending institution usually invests cash collateral in short term, high-quality investments that provide adequate liquidity given the short-term nature of the loans. This is the typical way for the process to work. The risks inherent in these transactions are not very different from any other investments, except that there is an added layer of leverage (i.e., that finances the cash collateral investments). Because the borrowers can often request the cash back at any time, it is prudent for the lender to ensure that the cash collateral investments provide adequate liquidity. Securities lending, therefore, is simply a leveraging “game” with the lion’s share of the risk borne by the lender, and largely a function of how the cash collateral is invested. The risks to the borrower are minimal, as the lender has an unconditional obligation to return the cash collateral at the termination of the loan. Additionally, because the borrower holds the securities of the lending institution as collateral, and has the contractual right to liquidate those securities in the event of a lender default, there is little, if any, systemic risk associated with a lender’s insolvency.

19. This writing is based on discussions with experts in the field and study, along with examination of the 2010 New York state insurance regulation for this activity. © 2012 National Association of Insurance Commissioners

258

Journal of Insurance Regulation

Thus, securities lending is an activity whereby an investor (it can be an insurer) loans its securities to a borrower, typically a large brokerage firm, for cash collateral.20 The industry practice requires cash collateral to be marked to market daily to ensure the value of the cash pledged by the borrower exceeds the value of the securities loaned.21 The lender pays the borrower a financing rate for the use of the pledged cash collateral and invests the cash collateral in anticipation of earning a return higher than the financing rate. Most securities loans are not only shortterm, but terminable on demand by either party. At the time of the AIG collapse, the New York state insurance regulators required cash collateral equal to 102 percent of the value of the loaned securities. This regulation was enforced and, if the lender did not follow it, the lender had to hold additional cash collateral for such a breach. Reading the 2008 10K of AIG and talking to experts, the securities lending activities at AIG were not typical for insurers or others in the securities lending industry for several reasons. One reason was that AIG did not maintain the collateral at the required 102 percent. This required AIG to have large cash collateral pools set aside to support any shortfall. During the crisis, and specifically after the announcement of the Lehman bankruptcy, a reduction in the market value of loaned securities, coupled with less demand to borrow securities, forced many lenders to make additional loans to maintain enough funding to support their cash collateral investment pools. The alternative would have been to sell assets that had been purchased with cash collateral and potentially realize losses.22 Throughout this period, most lenders did not realize losses and were able to maintain the necessary funding to support their investment pools. AIG, however, ran a significantly larger and more aggressive securities lending program than others. First, although their financing was generally short-term (typically 30 days or less), a large percentage of the cash collateral was invested in long-term fixed-income securities, with much of it in relatively illiquid MBS. Second, AIG relied heavily on corporate bonds, not just U.S. Treasuries and other “no risk” securities to fund their cash collateral investments. The demand by borrowers for U.S. Treasury and Agency collateral has always been quite strong, and this was true during the crisis, as investors flocked to low-risk and/or no-risk investments.23 As such, it has been common practice for lenders to maintain a high percentage, often up to 100 percent, of their 20. Other forms of collateral may be pledged, such as U.S. Treasury securities, but, for the purpose of this discussion, it is assumed that cash is pledged. 21. The NAIC Investments of Insurers Model Act (#280) requires that collateral equals 102 percent of the value of the loaned securities and be maintained at that level. As of 2010, the New York state insurance regulators added much stronger requirements for securities lending reporting, per Circular Letter 16 (2010). 22. Although cash collateral vehicles typically maintain a large percentage of investments in highly liquid assets, the unique circumstances during the crisis caught many securities lenders off guard. 23. Securities borrowed from lenders, particularly U.S. government and Agency securities, are often rehypothecated to money market investors under repurchase agreements. © 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

259

U.S. Treasury and Agency securities on loan. This proved to be the case during the crisis. The demand for corporate bonds, and most other securities, however, has been largely limited to specific issues that must be borrowed to cover a short. Therefore, the percentage of a lender’s corporate bonds placed on loan has been typically low, usually in the mid-to-low single digits.24 Unlike other lenders, AIG financed a large percentage of their cash collateral investments by “pushing out” corporate bonds to borrowers by paying relatively aggressive rates. This strategy provided the necessary financing at a relatively low premium until the crisis began to unfold. With AIG’s downgrades and housing market exposure, and a more acute realization of how heavily AIG was dependent on rolling their corporate bond loans, AIG’s borrowers began to increase the rate premium and demand additional collateral.25 As the crisis continued, the financing cost under this AIG strategy became extremely expensive. AIG eventually chose to terminate the loans and return the cash collateral to the borrowers, which they funded by loaning securities vs. cash collateral to the Federal Reserve Bank of New York.26 This action effectively replaced the funding provided by AIG’s securities lending borrowers with cash collateral provided by the Fed. This move was not considered necessary to avoid a collapse of AIG; rather, it was a strategy by AIG to stabilize its borrowings and the associated costs supporting their cash collateral investments. There was no systemic risk in this activity because the borrowers, in the event AIG defaulted on the loan transactions, had a contractual right to liquidate the securities borrowed from AIG and terminate the loans. And, the additional collateral pledged by AIG provided protection to the borrowers in the event the market value of the borrowed securities was less than the cash collateral pledged to AIG. The assertion that the securities lending activity would on its own have created the failure of AIG and a systemic collapse is in question after understanding the nature of the securities lending activity, as a default by AIG would not have directly caused harm to the borrowers. AIG’s aggressive securities lending activity required cash. It appears that the activity continued as if there were no liquidity crisis at AIGFP for posting cash for the CDS. It is apparent that there were breakdowns in the company’s lines of communication and no risk management. The explanation here serves to show that it was AIG that made the choice to close the loans and take $43 billion in bailout for the securities lending activity. In the spiral of AIGFP’s collapse, many activities that could have sustained themselves were caught in the turmoil. The securities lending activity is such a case in point. 24. According to the Risk Management Agency (RMA) quarterly surveys, the average amount of corporate bonds on loan during 2010 and 2009 was 5 percent and 3.5 percent, respectively. 25. As noted, market practice requires the borrower to pledge to the lender 120 percent of the value of the borrowed securities. With AIG, however, it is understood that most borrowers required AIG to post additional collateral (i.e., beyond the market value of the loaned securities), and many required that the total collateral held exceed 120 percent of the cash loaned. 26. According to a Federal Reserve press release dated Oct. 8, 2008, the Federal Reserve Board authorized the Federal Reserve Bank of New York to borrow up to$38.7 billion in fixedincome securities from AIG in return for cash collateral. © 2012 National Association of Insurance Commissioners

260

Journal of Insurance Regulation

Insurance Regulation The paper “Trends in Insurance Regulation”27 provides a history of insurance regulation in the United States during the past 200 years. The study shows that the insurance regulatory system in the United States is made up of a comprehensive body of insurance codes, laws and regulations that are built on experiences, past disasters and response to troubles. It is an ever-evolving and dynamic body of work. With each disaster, important new laws and regulations are enacted to protect the public. The U.S. insurance regulatory system can serve as a statistically significant sample of insurance regulation in the developed nations worldwide. It is a strong system that minimizes regulatory gaps with capital requirements, reserving asset allocation, contracts and market conduct oversight regulations. The state-based regulators of insurers and insurance prod ucts exert strong oversight in regard to the contracts, the liabilities and assets.28/29 With this strong body of laws and regulations, the foundation exists to ensure that if an insurer becomes insolvent, there is an orderly process of resolution that minimizes the harm to consumers. The regulatory mechanism has built-in processes to catch potential trouble early on. Statistical analysis is continually being conducted to find triggers for spot examinations. The NAIC created a comprehensive early warning system and many of the U.S. states have created their own, such as Texas. Academicians and actuaries have provided countless studies to supplement and provide a strong foundation for the regulators.30 It is well-known that insurers are overcapitalized relative to the requirements of the RBC laws. Moreover, each new product, such as variable annuities with guarantees, requires additional reserves (as well as additional capital) under the RBC formulas. In light of the strong insurance regulatory oversight in the United States, it is inconceivable that insurers, under their own watchdog, could have acted as AIGFP did with the massive CDS exposure. If by chance anyone had considered CDS to be an insurance product, how is it conceivable that it would have been allowed to not go through the rigorous insurance regulatory tests for sustainability, accurate actuarial assumption behind the rates, correct wording in the contracts and some built-in safety valves? The notoriously “burdensome” insurance regulation arm should not be shunned or ignored when the macroprudential regulations are set up 27. Baranoff and Baranoff (2003). 28. For a detailed explanation of insurance regulation, see the NAIC website (www.naic.org) and myriad textbooks. 29. Additional explanation of the regulatory structure of insurance is provided in Baranoff et al. (2009), Chapter 8. Chapter 9 and Chapter 10 provide an in-depth overview of the structure of insurance contracts. The textbook also provide insurance policies as examples of the insurance contracts allowed by law in the United States. 30. Many such studies appear in the Journal of Risk and Insurance, the two academic journals of The Geneva Association (The Geneva Papers on Risk and Insurance and The Geneva Risk and Insurance Review), the Journal of Insurance Regulation and many more. In order to avoid missing any work in the field of solvency detection and capital structure for insurers, the authors are not noted by name. The reader is invited to search these journals. © 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

261

to identify SIFIs. AIG did not fail under U.S. state-based insurance regulation. It can be said that not having this level of scrutiny was a macro factor that led to AIG’s collapse, as explored in more detail in the next parts of this report. Thus, in summary, the inside activities of AIGFP did fit well within the current definitions used by The Geneva Association’s SRI reports (2010a and 2010b): size, interconnectedness, substitutability and timing. AIGFP, a noninsurance subsidiary of AIG, adopted activities that used the size, stature and reputation of its holding company to create CDS. The CDS were interconnected to the global economies, and AIG did not have any substitutions and did not have enough time to raise cash when the liquidity disaster hit. In addition, the external (macro) underpinning factors that lead to systemic risks are as follows: the market conditions (housing markets bubble, subprime loans and entitlement ideology); regulatory gaps; and dependency on credit ratings instead of capital. We have indicated that CDS are not insurance contracts. It is important to point out that the securities lending activities and investment in MBS would not have caused the collapse on their own, as they did not take any other insurer down. In addition, all of AIG’s insurance subsidiaries’ policyholder claims were paid.

IV. Inside Factors: Innovation in Financial Products—CDS Much was written about the reasons for the near collapse of AIG and the subsequent bailout. As noted above—according to Harrington (2009), Sjostrum (2009) and myriad papers exploring the topic—AIG failed because it lacked the liquidity needed to post collateral for its CDS sold under AIGFP. As commonly noted, the need for cash for the securities lending activities, and the massive investments in MBS aggravated the situation. We explained above that the key factor was the CDS, while the latter two factors became part of the story as AIG was caught with liquidity needs in the spiral of the demise. The sale of the CDS was done through the AIGFP unit, established in 1987 under the AIG holding company, a non-insurance holding company operating under a thrift license from the U.S. government. AIGFP was described as an enterprise that “evolved into an indispensable aid to such investment banks as Goldman Sachs and Merrill Lynch, as well as governments, municipalities and corporations around the world. The firm developed innovative solutions for its clients, including new methods to free up cash, get rid of debt and guard against rising interest rates or currency fluctuations.”31 Initially, AIGFP prized itself for its careful financial modelers and risk engineers. Every model was examined and scrutinized. “Scepticism was hardwired into the company’s culture, part of its mantra: Hedge if you can. Don’t make speculative trades… assessing data daily, recalibrating assumptions constantly, 31. O’Harrow Jr. and Dennis (2008). © 2012 National Association of Insurance Commissioners

262

Journal of Insurance Regulation

counterbalancing one risk against another and making the hedges.”32 However, in 1998, the unit began selling CDS (unregulated derivative instruments) and, as a consequence, the unit became increasingly dependent on the success of these kinds of transactions. The models for CDS that had been adopted by AIGFP showed almost no likelihood of losing money. It was recognized as “free money” (for fees) as long as AIG kept its high rating.33 AIGFP thought it only covered the highest quality securitized bundles of debt and mortgages. As such, they did not see a reason to hedge the risk of providing “an apparent insurance” to the activities of collateralized debt obligations (CDOs) and MBS. In 2005, the unit discovered that the credit quality of the CDOs and MBS was much lower than assumed. AIGFP could not get out of its obligations, as is gleaned from the testimony of Joseph J. Cassano, former president of AIGFP, before the Financial Crisis Inquiry Commission (June 30, 2010.) What made it worse for the AIG holding company was the fact that there were apparent breakdowns in its internal risk management. With regard to the two factors noted above, the securities lending and the investments in MBS were part of the insurance operations structure of AIG and under the insurance regulatory umbrella. These activities, as we noted earlier, would not on their own have caused the collapse. Many insurers invested in MBS and some were involved in securities lending. However, except for The Hartford and Lincoln National, no other insurer resorted to accepting a bailout from the U.S. government. For these last two insurers, the amount received was a small fraction relative to the amounts received by banks and AIG (see Harrington, 2009). The insurance operations and structure of AIG were solid during the crisis, as was the case with most of the insurance industry. It has been well-established that AIG policyholders did not lose money, nor were they denied claims payments. Despite its liquidity needs, AIG’s insurance structure was solid. It was only because of the CDS exposure that liquidity was drained from AIG. In absence of the CDS collapse, the other two activities discussed in the literature would not have caused the collapse of AIG. Thus, the highly rated global AIG conglomerate incorporated in its structure (as shown in Figure 1, Figure 2 and Figure 4) a financial products unit that used the most sophisticated financial engineering modeling and products by talented financial architects to enhance profitability. As will be explained in the next section about the external macro factors that contributed to AIG’s failure, the accuracy of the assumptions behind the models and the triple-A rating were keys to the success of the unit. The innovation and the strength of the CDS were only as good as the quality of the MBS and the ability of AIG to keep its high rating. As the subprime mortgage market began to buckle, the models collapsed and with them the high rating of AIG. As the CDS contracts called for posting of cash

32. See footnote 26. 3. It is a paradox that sophisticated individuals would truly believe that they could get “free money,” especially in the corridors of such a massive and highly rated company as AIG. © 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

263

collateral in the case of rating downgrades, the liquidity crunch erupted. In the following section, we focus on the external macro factors. Thus, the key internal (inside) factors to the AIG collapse (not in insurance operations) were: 1) dependency on the success of the insurance operations to get high credit ratings; 2) use of regulatory arbitrage to build innovative and unregulated financial products; 30 faulty sophisticated financial models that included optimistic ratings assumptions and inaccuracies regarding the quality of underlying bundled securities; 4) complex and unsustainable CDS contract terms; 5) the CDS growth; 6) breakdown in internal risk management; and 7) deficient internal controls. Internal factors (in insurance operations) that were caught in the liquidity crisis spiral were the securities lending and investments in MBS.

V.

External Macro Factors: Housing Boom Driven by Lenient Underwriting of Mortgages; Bundling of Debt Instruments; Permissive Banking/Thrift Regulation; No Derivative Regulations—Incentives to the Explosion in CDS Market

Insurance regulation has a proven track record of having avoided any failure of an insurance company with systemic consequences. AIGFP was not regulated by insurance regulators as shown in Figure 1, Figure 3 and Figure 4. AIGFP was a subsidiary of the AIG holding company, which, in turn, was supervised by the OTS.34 In fact, AIGFP began operating out of London in 1987 and was regulated by French banking regulators.35 The Financial Crisis Inquiry Report (National Commission on the Causes of the Financial and Economic Crisis in the United States, 2011) notes that: 34. See Harrington (2009), p. 799 “The assertion that AIGFP was unregulated is technically incorrect and appears misleading. As noted above, and as a consequence of owning a savings and loan subsidiary, AIG was subject to consolidated regulation and oversight by the OTS, and it was recognized as such for the purpose of meeting the 2005 E.U. regulatory criterion for group supervision.” 35. Based on AIG 2009 10 K: “A total of $234.4 billion (consisting of corporate loans and prime residential mortgages) in net notional exposure of AIGFP’s super senior credit default swap portfolio as of December 31, 2008 represented derivatives written for financial institutions, principally in Europe, for the purpose of providing regulatory capital relief rather than for arbitrage purposes. These transactions were entered into by Banque AIG, AIGFP’s French regulated bank subsidiary, and written on diversified pools of residential mortgages and corporate loans (made to both large corporations and small to medium sized enterprises). In exchange for a periodic fee, the counterparties receive credit protection with respect to diversified loan portfolios they own, thus reducing their minimum capital requirements.” See http://washingtonoutside. blogspot.com/2010_03_01_archive.html. © 2012 National Association of Insurance Commissioners

264

Journal of Insurance Regulation

The Office of Thrift Supervision has acknowledged failures in its oversight of AIG (holding company)….John Reich, a former OTS director, told the FCIC that… he had “no clue—no idea—what [AIG’s] CDS liability was…. the OTS’s authority to regulate holding companies was intended to ensure the safety and soundness of the FDIC-insured subsidiary of AIG and not to focus on the potential impact on AIG of an uninsured subsidiary like AIG Financial Products…. Finn ignored the OTS’s responsibilities under the European Union’s Financial Conglomerates Directive (FCD)—responsibilities the OTS had actively sought. The directive required foreign companies doing business in Europe to have the equivalent of a “consolidated supervisor” in their home country…..Reich told FCIC staff that he did not understand his agency’s responsibilities under the FCD….The OTS did not look carefully at the credit default swap portfolio guaranteed by the parent company—even though AIG did describe the nature of its super-senior portfolio in its annual reports at that time, including the dollar amount of total credit default swaps that it had written….. In February 2008, AIG reported billions of dollars in losses and material weaknesses in the way it valued credit default swap positions. Yet the OTS did not initiate an in-depth review of the credit default swaps until September 2008—ten days before AIG went to the Fed seeking a rescue….He (Reich) also acknowledged that the OTS had never fully understood the Financial Products unit, and thus couldn’t regulate it…. Reich said that for the OTS to think it could regulate AIG was “totally impractical and unrealistic. AIGFP did not have any strong oversight. In the United States, insurers are not permitted to sell CDS and, as mentioned, CDS were not considered to be insurance (see Stulz, 2010); therefore, they were not regulated by state insurance regulators. This led to a gap in controls over those products and the actions regarding such products by AIGFP. This regulatory gap was clearly used to develop the innovative financial products such as CDS by AIGFP. The permissive regulatory framework for banks and thrifts, combined with no regulation of derivatives under the “free markets” philosophy (see PBS’ “Frontline” expose;36 Harrington, 200937 and Levine, 201038) should be regarded as major propelling

36. PBS “Frontline” at www.pbs.org/wgbh/pages/frontline/warning. 37. See Harrington (2009), p. 800: “Banking regulation permitted and probably encouraged high leverage, aggressive investment strategies, inadequate capital requirements for risky loans and securitizations, and complex off-balance sheet vehicles, often financed by commercial paper, all taking place within the framework of government deposit insurance and “too big to fail” (TBTF) policy.” © 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

265

factors to the expansion of the CDS market. Because banks could substitute capital by buying CDS from highly rated institutions, the product took off and AIG was a major player with huge exposure in its AIGFP unit. The huge CDS exposure of the AIGFP unit was fueled by the housing boom, subprime mortgages and securitization activities. According to O’Harrow and Dennis, “By 2004, Wall Street investment banks were discovering how to turn consumer debt into a moneymaker, churning out bond-like securities backed by mortgages and other assets.”39 CDS gave the hedging necessary to attract institutional investors to the CDOs and MBS, as seen in Figure 2 and Figure 3. The ideology that “everyone deserves to own a home,” along with the allowances for greater leveraging, using securitization, fueled a spiral of need for loans that would be later bundled. The subprime mortgages that emerged by loaning money to unqualified home buyers penetrated even the highest credit layers of bundled securities. The idea that home prices would continue to escalate and unqualified borrowers would be able to sell their home at a profit when they could no longer pay the higher variable mortgage payments fell apart. In retrospect, federal regulators and the U.S. Securities and Exchange Commission (SEC) did admit (see PBS’ “Frontline” expose) their major mistake in not regulating the derivatives markets and products. It is unequivocally recognized that this gap in regulation was a key contributor to the financial crisis of 2007–2009. Thus, key external macro factors to the AIG collapse were: 1) the free markets philosophy of self-discipline; 2) the “Everyone deserves to own a home” ideology, in addition to a housing boom with lenient underwriting standards and subprime mortgages; 3) dependency on credit rating agencies; 4) permissive banking and thrift regulation (i.e., financial products such as CDS of highly rated institutions serving as substitutes for capital); 5) no derivatives (CDS) regulation (i.e., no checks and balances regarding the CDS market); 6) growth of the bundling of securities with underlying lax underwriting standards for borrowers; and 7) insurance regulators not allowed in despite of the nature of CDS contracts as providing some “supposed” security.

38. Levine (2010), p. 197. Levine studies the following: 1) SEC policies toward credit rating agencies; 2) Federal Reserve policies that allowed banks to reduce their capital cushions through the use of CDS; 3) SEC and Federal Reserve policies concerning over-the-counter (OTC) derivatives; 4) SEC policies toward the consolidated supervision of major investment banks; and 5) federal government policies toward two housing-finance entities, Fannie Mae and Freddie Mac. 39. O’Harrow Jr. and Dennis (2008). © 2012 National Association of Insurance Commissioners

266

Journal of Insurance Regulation

VI. AIG’s Inside Factors and External Macro Factors Come Together: AIGFP Unit Unable to Untangle Itself from the CDS Obligations when the Housing Markets Collapsed While the insurance operations of AIG that generated the high credit ratings was under strong insurance regulatory scrutiny, AIGFP was under the lenient regulatory mechanism described above. In the collision between the internal and external factors, it is key to note that the part of AIG holding company that was not under the insurance regulatory mechanism, is the part that led to the demise (see Figure 4). The underpinning macro factors (that did not encompass the AIG insurance subsidiaries) led to creating the CDS, the risky financial products. Without controls, these non-insurance products under non-insurance structure and regulation led to the demise. Following is a quick description of how the factors aggravated each other to the point of destruction. In 2005, the AIGFP unit discovered that the credit quality of the debt obligations it secured was much lower than assumed. When it could not reverse the obligations, it aggravated the situation by keeping this information in house (at the AIGFP unit) without transparency. The internal controls were shot and the unit did not disclose its troubles. As the situation deteriorated, the housing bust led to credit rating downgrades and the spiral of the liquidity crisis erupted. The set-up of the CDS contracts did not provide for gradual cash settlement, but rather required major cash collateral upon a downgrading event. If AIG had added “installments” into its CDS contracts, the gradual needs might have prevented the sudden placement of large cash collaterals, as discussed in the appendix of “what ifs?” But that was not the case. Thus, the combination of no internal risk management and no external regulatory requirements and controls led to lack of transparency regarding the size of AIG’s CDS exposure. The size of the CDS exposure was allowed to mushroom. As banks needed more and more CDS to replace capital (which was permitted by lenient banking regulation) and more bundling occurred in the MBS, the CDS market boomed. The housing market crash led to AIG’s calamity without an apparent way out. AIGFP could not unwind its major exposure and could not stop the AIG rating downgrades. Each downgrade led to calls for posting cash collateral.40 The liquidity position of AIG deteriorated fast as the rating agencies continued to downgrade this large global financial conglomerate. This spiral led to the bailout. On the way, the securities lending activity and the investments in MBS were caught in the spiral and required more cash, as well. This exacerbated the liquidity drain. Interestingly, in this collision of factors, the story has nothing to do with AIG being an insurer selling insurance products. Without the CDS, even if the 40. Note the terminology. The banks did not submit “claims” on policies, as the term is used and understood in insurance terminology. © 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

267

securities lending and the investments in the MBS had been the only triggers to the enormous problem for AIG, the regular insurance resolution process of insolvencies would not have led to a systemic collapse in need of bailout. The collision of factors had nothing to do with the financial behavior of insurers, their underwriting or their asset allocation activities.41 CDS could have been sold by any player with a high credit rating, reputation and standing. The product was apparently sold by other insurers, but there was no such large exposure and there was more careful treatment.42 AIG’s high credit rating was a prerequisite to such sales and acceptance by banks and their regulators.43 The AIGFP unit was unique in its use of every means of regulatory arbitrage as an opportunity and belief in its sophisticated models and its own wisdom to a fault. Thus, the key external macro and activity-related factors that came together and led to the AIG collapse were: 1) housing market collapsed; 2) AIG credit rating fell; 3) cash collateral calls under the CDS contracts; 4) liquidity crisis pervasive; and 5) no time to raise funds and the federal government bails out.

VII. Summary and Conclusions The AIG failure was caused by a mix of inside factors and external macro factors. The inside factors included: reliance on the strength of the AIG insurance operations to obtain a high credit rating for innovation in financial products outside of the insurance operations; CDS contract design; faulty financial models; and lack of appropriate risk management. The external macro factors included: lax regulation of banks and thrifts, with no regulation of derivatives (free markets ideology); housing market bubble and collapse (entitlement-to-homeownership ideology); securitization growth; and reliance on credit rating agencies. All of these factors came to a head to bring about the collapse: no regulation of sophisticated financial products with high credit ratings => growth in the CDS market => faulty models not predicting housing markets’ collapse and rating downgrades => calls for cash collaterals => liquidity disaster => failure and bailout. The delineation of the factors contributing to the demise of AIG leads us to the following key conclusions and lessons in the area of systemic risk: 1.

If a high credit rating is allowed to replace capital, regulators need to understand the systemic implication of such decisions. a.

When innovative financial products use a credit rating established by strong insurance operations, regulators need to

41. For the interested reader, explanation is available in chapter 7 of Baranoff et al. (2009). 42. Prudential Plc of the U.K. owned Egg and sold it to CITI in 2006. Egg sold CDS and suffered losses in the U.K. due to the housing and credit markets there. See Prudential Plc 2006 10K. 43. Levine (2010), pp. 196–213. © 2012 National Association of Insurance Commissioners

Journal of Insurance Regulation

268

b. c. 2.

understand the implication of allowing this high rating to support banks’ capital. There needs to be recognition of the danger that the solidity and stability of insurance can provide. Can this solidity be sustained without some macroprudential actions? Regulators need to understand fully the merits behind models such as those created by AIGFP to substantiate selling so much CDS exposure. Replacement of capital by any other products needs complete transparency and understanding.

Regulatory gaps need to be closed to avoid arbitrage. The absence of derivative regulations was a major factor to the growth in the innovative derivative products. a. b.

No posting of liabilities was required. There were no “checks and balances,” nor was there an understanding of the risks.

3.

Insurance institutions and their operations should be required to adhere to insurance regulation. When insurers look into alternative regulatory frameworks, it could indicate a move into non-insurance or quasi-banking products. This may trigger new risks not known to insurance regulators.

4.

There should not be confusion between insurance and non-insurance (quasi-banking) activities.

The umbrella of insurance regulation has proven itself to be a source of stability. Insurance by its nature is the antidote to risk; i.e., a solution to mitigating risks. It is a propeller of the economy and, when it is done under the well-proven success of state-based insurance laws and regulations, it is a source of stability and a reason to “sleep well at night.” As such, insurance regulators should not shunned away from products or entities that emulate “insurance” without really being insurance.

© 2012 National Association of Insurance Commissioners

Analysis of the AIG Case

269

References AIG (2008) “Securities Lending Agreement with Federal Reserve Bank of New York, Oct. 8, accessible at www.aig.com/aigweb/internet/en/files/Securities %20LendingTalking%20Pts%20(FINAL)%20(2)_tcm20-125115.pdf. AIG 10K (2008) “Continuing Liquidity Pressures” p. 6 and p. 42. Accessible at: www.sec.gov/Archives/edgar/data/5272/000095012309003734/y74794e10vk.htm. AIG 10 K (2009). Baranoff Etti, Lee Brockett, Patrick and Kahane, Yehuda (2009) Risk Management for Enterprises and Individuals, Flat World Knowledge. Baranoff, Etti G. and Baranoff, Dalit (2003) “Trends in Insurance Regulation” Review of Business, Fall, pp. 11–20. Desmond, Maurna (2008) “AIG’s Play For Time,” Forbes.com, accessible at www.forbes.com/2008/10/12/aig-federal-reserve-markets-equitycx_md_ 1010markets35.html. Harrington, S.E., (2009) “The Financial Crisis, Systemic Risk, and the Future of Insurance Regulation,” Journal of Risk and Insurance 76, pp. 785–819. Federal Reserve (2008), Press Release, 8 October 2008. http://federalreserve.gov/ newsevents/press/other/20081008a.htm. Liedtke, Patrick (2010) “The Lack of an Appropriate definition of Systemic Risk” The Geneva Association Insurance and Finance Newsletter No. 6, July. Levine, Ross (2010) “An autopsy of the US financial system: accident, suicide, or negligent homicide” Journal of Financial Economic Policy, Vol. 2, No. 3, pp. 196–213. National Commission on the Causes of the Financial Crisis in the United States (2011) The Financial Crisis Inquiry Report, Released in January 2011. O'Harrow Jr., Robert. and Dennis, Brady (2008) “The Beautiful Machine,” Washington Post Staff Writers, The Washington Post, three parts beginning Dec. 29, accessible at www.washingtonpost.com/wp-dyn/content/article/2008/ 12/28/AR2008122801916.html?hpid=topnews&s_pos=. Orol, Ronald D. (2010) “Geithner, Paulson defend $182 billion AIG bailout: Lawmakers grill Geithner and Paulson for failing to obtain taxpayer concessions”, MarketWatch, Wall Street Journal, Jan. 27, accessible at www.marketwatch.com/story/geithner-paulson-defend-182-bln-aig-bailout2010-01-27. PBS “Frontline” (2009) “The Warning” accessible at www.pbs.org/wgbh/pages/ frontline/warning. Prudential Plc 10K (2006) http://quote.morningstar.com/stock-filing/AnnualReport/2006/ 12/31/t.aspx?t=XNYS:PUK&ft=&d=8106c331f01bfa4b. Shelp, Ron (2010) “Shelp Interview About AIG, Cassano,” Bloomberg, www.bloomberg.com/ video/61196958/. Shelp, Ron (2006) Fallen Giant: The Amazing Story of Hank Greenberg and the History of AIG John Wiley & Sons, Hoboken, N.J.

© 2012 National Association of Insurance Commissioners

270

Journal of Insurance Regulation

Sjostrom, W.K. Jr. (2009) “The AIG Bailout,” Washington & Lee Law Review, pp. 943–991. Accessible at http://ssrn.com/abstract=1346552. Sorkin, Andrew Ross (2009) “The Case for Saving AIG, by AIG” New York Times, March 2, accessible at www.nytimes.com/2009/03/03/ business/ economy/03sorkin.html?_r=1. Stulz, René M. (2010) “Credit Default Swaps and the Credit Crisis,” Journal of Economic Perspectives—Volume 24, Number 1, Winter, pp.73–92. TradingMarkets.com (2011) “Federal Commission: AIG Meltdown Exemplified ‘Breakdowns and Irresponsibility,” Jan. 27 (BestWire Services, 2011). Accessible at www.tradingmarkets.com/news/stock-alert/aig_federalcommission-aig-meltdown-exemplified-breakdowns-and-irresponsibility-1451012.html. The Geneva Association (2011) Insurance and Finance Newsletter No. 7, February, accessible at www.genevaassociation.org/PDF/Insurance_And_ Finance/GA2011-I&F07.pdf. The Geneva Association (2010a) Systemic Risk in Insurance—An analysis of insurance and financial stability, The Geneva Association, Geneva, March, accessible at www.genevaassociation.org/PDF/BookandMonographs/Geneva_ Association_Systemic_Risk_in_Insurance_Report_March2010.pdf. The Geneva Association (2010b) Key Financial Stability Issues in Insurance—An account of The Geneva Association ongoing dialogue on systemic risk with regulators and policy-makers, The Geneva Association, Geneva, July, accessible at www.genevaassociation.org/PDF/BookandMonographs/Geneva_Association_ Key_Financial_Stability_Issues_in_Insurance_July2010.pdf.

© 2012 National Association of Insurance Commissioners

Copyright of Journal of Insurance Regulation is the property of National Association of Insurance Commissioners and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.